1. Field of the Invention
The present invention relates to a lines and characters separation apparatus for separating lines or characters from each other in a document page, the separation being one step in pattern recognition.
2. Description of the Prior Art
Several conventional lines and characters separation apparatuses have been proposed.
Examples of these apparatuses are an apparatus for scanning a document page to directly read line spacing according to density distributions of accumulation signals of character portions in lines and accumulation signals of non-character portions, and an apparatus for statistically calculating line spacing on the basis of the density distributions of the accumulation signals.
For each separated line, characters belonging to it are separated in the same manner as in line separation.
Raster or parallel scanning along a predetermined direction is performed as document page scanning. According to each scanning scheme, scanning time-serially progresses from a start point to an end point of the document page. The predetermined scanning direction is corrected by scanning the same document page along a corrected direction or by analyzing the previous data, i.e., by utilizing a processor.
In either above-described apparatus, the document page is time-serially scanned to require a long period of time for separating the lines and the characters.
Furthermore, correction of the scanning direction, which is required by a line inclination or the like, can be performed only after at least one scanning cycle is completed. Therefore, in this case, a longer period of time is required to separate the lines and the characters.